1. Field of the Invention
The present invention relates generally to an observation method and an observation apparatus for radio waves and acoustic waves, and more specifically, to an observation method and an observation apparatus which enables an observation without dead angle.
Also, the present invention relates to a multi-dimensional hologram data processing apparatus and a method for extracting a plurality of peak points for multi-dimensional hologram data and an area occupied thereby using the multidimensional hologram data processing apparatus.
2. Description of the Related Art
Since radio waves and acoustic waves are similarly waves, holograms of those waves can be observed as is the case with light and are utilized to visualize a wave source image and to specify noise (such as undesired electromagnetic wave radiation and ambient noise) sources. The inventors have disclosed a method and an apparatus for observing a radio wave hologram and acoustic wave hologram to derive a wave field intensity and wave source image, for example, in Japanese Laid-open Patent Application No. 8-201459 and Japanese Laid-open Patent Application No. 9-134113.
FIG. 1 is a schematic diagram for illustrating a hologram observation method described in Japanese Laid-open Patent Application No. 8-201459. Rectangular hologram observation surface 92 is set away from observation object (wave source) 91. Scanning sensor 93 which moves two-dimensionally in hologram observation surface 92 is used to detect radio waves and acoustic waves at a predetermined observation frequency from observation object 91 at each point in hologram observation surface 92. In addition, fixed sensor 94 is provided separately from scanning sensor 93 and is used to similarly detect radio waves and acoustic waves at the above-mentioned predetermined frequency from observation object 91. Signals from both sensors 93 and 94 are interfered with at interference unit 95 and the signals after interference is detected by detector 96. The detected signal (signal representing the correlation of the signals from both sensors 93 and 94, i.e. signal representing the hologram intensity at a position of scanning sensor 93 in hologram observation surface 92) is stored in memory 97 corresponding to the coordinates of scanning sensor 93 in hologram observation surface 92. When observations are completed at all observation points in hologram observation surface 92, data is read from memory 97 to reconstruct a hologram image by image reconstructing unit 98.
In the prior art hologram observation method as described above, however, plane scanning in the hologram observation surface is used to observe holograms, so that it is not possible to observe radio waves and acoustic waves arriving from the backside of the hologram observation surface. Additionally, it is difficult to observe radio waves and acoustic waves of extremely oblique incident angle with respect to the hologram observation surface. Thus, in reality, a field of view angle is as small as 120 degrees and the remaining angle distance of 240 degrees is a dead angle, which causes a disadvantage of a limited observation. A required observation can be made even with a relatively small field of view angle as described above, for example, when an observation object is placed at a corner of a room such as radio wave darkrooms and a hologram observation apparatus is placed at a corner opposite to the observation object. However, when an observation is made outdoors, radio waves and acoustic waves to be observed can not arrive only in a front direction. For this reason, many components of radio waves and acoustic waves, arriving other than in the front direction, are left without being observed, thus producing some space which can not be observed.
Furthermore, the prior art hologram observation method has a disadvantage of lacking a real-time basis observation since the hologram image is reconstructd after the data at all the observation points is acquired in the hologram observation surface.
Also, there are a circumference scanning type hologram observation for extracting each arrival angle (xcex8, xcfx86xe2x80x2) of a plurality of waves and a plane scanning type hologram observation for extracting each of the coordinates (Xs,Ys,Zs) of a plurality of points wave sources illustrated in FIG. 2 (Hitoshi Kitayoshi: xe2x80x9cStudy for visualizing electromagnetic radiation and propagationxe2x80x9d, second chapter xe2x80x9cprinciple of visualization and reconstruction algorithmxe2x80x9d in doctoral dissertation in Tohoku University, Feb. 1997).
Hologram observation data has an accuracy equal to or greater than the observed dimensions, for example, a three-dimensional image can be reconstructd from data recorded in a two-dimensional plane. However, the reconstructd image has a limited resolution due to a limitation of an observation surface as described in the above-mentioned literature for the plane scanning type hologram observation (Hitoshi Kitayoshi: xe2x80x9cStudy for visualizing electromagnetic radiation and propagationxe2x80x9d, second chapter xe2x80x9cprinciple of visualization and reconstruction algorithmxe2x80x9d in doctoral dissertation in Tohoku University, Febuary 1997, pp.13-19). Thus, when a plurality of wave sources are simultaneously observed, it is substantially difficult to automatically extract the position and intensity of each wave source.
Conventionally, a contour line processing method and a path survey method are used to detect peaks and an area occupied by the peaks as shown in FIG. 3. The path survey method is one for surveying the negative inclination path from a peak point in all moving directions to determine an area occupied by the point.
The above-mentioned prior art has disadvantages as described below.
Specifically, in the above-mentioned algorithm, the creation of a path for survey is complicated and is not easily implemented by simple hardware or digital signal processing (DSP).
Although another approach is also contemplated in which a reconstructd image is improved in extended peak (blurred image) by modifying a reconstructing algorithm for hologram images, the approach is not complete in the relationship between parameters used when applying the algorithm and the stability of the reconstructd image (Hitoshi Kitayoshi: xe2x80x9cStudy for visualizing electromagnetic radiation and propagationxe2x80x9d, second chapter xe2x80x9cprinciple of visualization and reconstruction algorithmxe2x80x9d in doctoral dissertation in Tohoku University, Febuary 1997, pp.20-35). The parameter refers to a threshold value in SPIM (Spectrum Phase Interpolation Method) (Hitoshi Kitayoshi: xe2x80x9chigher resolution for short time frequency spectrum analysisxe2x80x9d Shingakuron A, vol.J76-A, no.1, pp.78-81, January 1993., Hitoshi Kitayoshi: xe2x80x9chigher resolution for two-dimensional complex spectrum analysisxe2x80x9d Shingakuron A, vol.J76-A, no.4, pp.687-689, April 1933), while the parameter refers to a filter terms number and the like in MEM (Maximum Entropy Method) (Mikio Hino: xe2x80x9cSpectrum analysisxe2x80x9d Asakura Syoten, 1977, Yoshinao Aoki: xe2x80x9cWave signal processingxe2x80x9d Morikita Pub., sixth chapter xe2x80x9cMaximum Entropy Methodxe2x80x9d, 1986).
In view of the above-mentioned disadvantages in the prior art, it is an object of the present invention to provide a hologram observation method and apparatus which can have 360 degrees of a viewing angle without dead angle and make a real-time evaluation for the propagation of radio waves and acoustic waves.
It is another object of the present invention to provide a multi-dimensional hologram data processing apparatus which can be implemented by a hardware or digital signal processing (DSP) with a simple algorithm and a method for extracting a plurality of peak points for multi-dimensional hologram data and an area occupied thereby using the multi-dimensional hologram data processing apparatus.
The present invention provides a hologram observation method for measuring radiation waves from an observation object to reconstruct a hologram, wherein while a first sensor scans on a circumference, a radiation wave is received by the first sensor to generate a first received signal, the radiation wave is received by a second sensor at a position not changed with respect to the center of a circle with the circumference to generate a second received signal, the first received signal is made to interfere with the second received signal to acquire an interference signal, the interference signal is detected to obtain measured data at each point on the circumference.
In the observation method according to the present invention, it is preferable to calculate evaluation function V(xcfx86xe2x80x2) based on measured data Ez(r, xcfx86) at each point in a range of half of the circumference with the following equation:       V    ⁢          (              φ        xe2x80x2            )        =            ∫                        -          π                /        2                              +          π                /        2              ⁢                  W        ⁢                  (          φ          )                    ⁢              ⅇ                              -            2                    ⁢          πj          ⁢                      xe2x80x83                    ⁢          r          ⁢                      xe2x80x83                    ⁢          sin          ⁢                      xe2x80x83                    ⁢          θ          ⁢                      xe2x80x83                    ⁢                                    cos              ⁢                              (                φ                )                                      /            λ                              ⁢                        E          z                ⁢                  (                      r            ,                          φ              +                              φ                xe2x80x2                                              )                    ⁢              ⅆ        φ            
to estimate orientation xcfx86xe2x80x2 of the radiation wave. In the equation 3, rotation angle xcfx86 represents a point on the circumference, r represents a radius of the circle, j represents an imaginary unit, xcfx80 represents ratio of circumference of circle to its diameter, xcfx86xe2x80x2 represents a rotation angle at a position of the center in the range of the half of the circumference, xcex represents a wavelength of the radiation wave, xcex8 represents an incident angle of radiation wave with respect to the central axis of the circle, and W(xcfx86) represents a predetermined weighting function. In this event, a peak in evaluation function V(xcfx86xe2x80x2) can be selected to calculate incident angle xcex8 from an optimal value for rxc2x7sin xcex8. Alternatively, evaluation function V(xcfx86xe2x80x2) may be calculated while a rotation axis direction of the circle is changed such that the incident angle is equal to 90 degrees.
In the hologram observation method according to the present invention, it is preferable that the acquisition of the measured data is continuously executed by continuously moving the first sensor on the circumference and evaluation function V(xcfx86xe2x80x2) is continuously calculated based on the measured data for the half of the circumference out of the previously obtained measured data. In such a case, it is desirable that the present rotation angle of the first sensor is not included in the angle range of the half of the circumference for calculating evaluation function V(xcfx86xe2x80x2). Also, an image in a direction of rotation angle xcfx86xe2x80x2 may be photographed and the photographed image is displayed together with a display representing the calculated evaluation function V(xcfx86xe2x80x2).
The present invention provides a hologram observation apparatus for observing a radiation wave from an observation object to reconstruct a hologram, and has a scanning sensor for receiving the radiation wave to generate a first received signal; a driving means for driving the scanning sensor to perform scanning on the circumference; a fixed sensor placed at a position not changed with respect to the center of the circle for receiving the radiation waves to generate a second received signal; an interference unit for interfering with the first received signal and the second received signal to output an interference signal; and a detector for detecting the interference signal to output measured data at each point on the circumference. In this observation apparatus, a trigger timing for measurement may be determined based on an ID signal extracted from the second received signal. Additionally, the observation apparatus may have a level detecting means for calculating an average signal level from the second received signal and a level calibration unit for calibrating levels of the measured data based on the average signal level.
Furthermore, the observation apparatus preferably have a ring data buffer memory written with the measured data into an address in accordance with the present rotation angle of the scanning sensor corresponding to the circumference. It is also preferable for the observation apparatus to include an evaluation value calculating unit for calculating evaluation function V(xcfx86 ) based on measured data Ez(r,xcfx86) at each point in the above-mentioned range of the half of the circumference stored in the ring data buffer memory with the following equation:       V    ⁢          (              φ        xe2x80x2            )        =            ∫                        -          π                /        2                              +          π                /        2              ⁢                  W        ⁢                  (          φ          )                    ⁢              ⅇ                              -            2                    ⁢          πj          ⁢                      xe2x80x83                    ⁢          r          ⁢                      xe2x80x83                    ⁢          sin          ⁢                      xe2x80x83                    ⁢          θ          ⁢                      xe2x80x83                    ⁢                                    cos              ⁢                              (                φ                )                                      /            λ                              ⁢                        E          z                ⁢                  (                      r            ,                          φ              +                              φ                xe2x80x2                                              )                    ⁢              ⅆ        φ            
where rotation angle xcfx86 represents a point on the circumference, r represents a radius of the circle, j represents an imaginary unit, xcfx80 represents ratio of circumference of circle to its diameter, xcfx86xe2x80x2 represents a rotation angle at a position at the center in the range of the half of the circumference, xcex represents a wavelength of the radiation wave, xcex8 represents an incident angle of the radiation wave with respect to a plane of the circumference, and W(xcfx86) represents a predetermined weighting function. In this event, it is preferable to provide an offset adding unit for adding an angle distance difference for the half of the circumference to the present rotation angle of the scanning sensor to generate rotation angle xcfx86xe2x80x2. Moreover, the observation apparatus preferably have a TV camera rotatively driven by driving means together with the scanning sensor while the TV camera maintains an angle distance difference for the half of the circumference with respect to the scanning sensor, and a display unit for making a display of a photographed image by the TV camera, a display based on the measured data, and a display based on the evaluation function.
The principle of the hologram observation according to the present invention will be hereinafter described. Three-dimensional xyz rectangular coordinates are set, in which and observation point P is arranged at distance r from origin O in xy plane. Observation point P can rotate around origin O and a rotation angle of observation point P measured with respect to x axis is represented by xcfx86. A zenith angle (incident angle) of wave source S measured with respect to z axis is represented by xcex8. Here, assuming that a plane wave from wave source S enters observation point P.
As shown in FIG. 4(a) and FIG. 4(b), a cylindrical coordinates representation of the Maxwell equation is applied to the plane wave traveling in a direction making angle xcex8 against z axis with y axis representing magnetic field H to derive a z axis component of an electric field, i.e. Ez component, as follows:                                           E            z                    ⁡                      (                          r              ,              φ                        )                          =                              A            o                    ⁢                                    ⅇ                              j                ⁢                                  xe2x80x83                                ⁢                kz                ⁢                                  xe2x80x83                                ⁢                cos                ⁢                                  xe2x80x83                                ⁢                θ                                      ·            sin                    ⁢                      xe2x80x83                    ⁢          θ          ⁢                                    ∑                              n                =                                  -                  ∞                                            ∞                        ⁢                                                            (                  j                  )                                n                            ⁢                                                J                  n                                ⁡                                  (                                      k                    ⁢                                          xe2x80x83                                        ⁢                    r                    ⁢                                          xe2x80x83                                        ⁢                    sin                    ⁢                                          xe2x80x83                                        ⁢                    θ                                    )                                            ⁢                              ⅇ                                  j                  ⁢                                      xe2x80x83                                    ⁢                  n                  ⁢                                      xe2x80x83                                    ⁢                  φ                                                                                        (        1        )            
where j represents the imaginary unit, Jn represents the Bessel function, and k=2 xcfx80/xcex, xcex represents a wavelength of the wave to be observed.
Expression (1) can be developed using the Jacobi development formula as follows:
Ez(r,xcfx86)=Aoejkz cos xcex8xc2x7sin xcex8ejkr sin xcex8 cos xcfx86xe2x80x83xe2x80x83(2)
Electric field Ez on the circumference when Z=0 is derived as follows:
Ez(r,xcfx86)=Aoxc2x7sin xcex8ejkr sin xcex8 cos xcfx86xe2x80x83xe2x80x83(3)
Assuming that an incident angle of the plane wave with respect to a direction of x axis is xcfx86i, expression (3) can be transformed as follows:
Ez(r,xcfx86)=Aoxc2x7sin xcex8ejkr sin xcex8 cos(xcfx86xe2x88x92xcfx86i)xe2x80x83xe2x80x83(4)
In expression (4), it is assumed that A0, xcex8, and xcfx86i are unknown and only electric field Ez(r, xcfx86) on the circumference at distance r from origin O can be observed in plane z=0 (xy plane). Here, evaluation function V(xcfx86xe2x80x2) is defined as follows:                               V          ⁡                      (                          φ              xe2x80x2                        )                          =                              ∫                                          -                π                            /              2                                      π              /              2                                ⁢                                    W              ⁡                              (                φ                )                                      ⁢                          ⅇ                                                -                  j                                ⁢                                  xe2x80x83                                ⁢                                  kr                  xe2x80x2                                ⁢                                  xe2x80x83                                ⁢                cos                ⁢                                  xe2x80x83                                ⁢                φ                                      ⁢                                          E                z                            ⁡                              (                                  r                  ,                                      φ                    +                                          φ                      xe2x80x2                                                                      )                                      ⁢                          ⅆ              φ                                                          (        5        )            
Here, W(xcfx86) is a weighting function for the purpose of stabilization (decreasing cut-off error) of evaluation function V(xcfx86xe2x80x2). For example, assuming
W(xcfx86)={fraction (1/xcfx80)}(1+cos(2xcfx86))xe2x80x83xe2x80x83(6)
, then                                           ∫                                          -                π                            /              2                                      π              /              2                                ⁢                                    W              ⁡                              (                φ                )                                      ⁢                          ⅆ              φ                                      =        1                            (        7        )            
is derived, i.e. it can be a standardized weighting function.
When expression (4) is substituted into expression (5), the following is derived.                               V          ⁡                      (                          φ              xe2x80x2                        )                          =                              ∫                                          -                π                            /              2                                      π              /              2                                ⁢                                    W              ⁡                              (                φ                )                                      ⁢                          ⅇ                                                -                  j                                ⁢                                  xe2x80x83                                ⁢                                  kr                  xe2x80x2                                ⁢                cos                ⁢                                  xe2x80x83                                ⁢                φ                                      ⁢                                          A                o                            ·              sin                        ⁢                          xe2x80x83                        ⁢            θ            ⁢                          xe2x80x83                        ⁢                          ⅇ                              j                ⁢                                  xe2x80x83                                ⁢                kr                ⁢                                  xe2x80x83                                ⁢                sin                ⁢                                  xe2x80x83                                ⁢                θ                ⁢                                  xe2x80x83                                ⁢                                  cos                  ⁡                                      (                                          φ                      +                                              φ                        xe2x80x2                                            -                                              φ                        i                                                              )                                                                        ⁢                          ⅆ              φ                                                          (        8        )            
In expression (8), assuming that weighting function W(xcfx86) is a standardized weighting function as shown in expression (6) and rxe2x80x2=rxc2x7sin xcex8, then the following is obtained when xcfx86xe2x80x2=xcfx86I;
V(xcfx86I)=A0xc2x7sin xcex8
Also, as is apparent from expression (8), V(xcfx86xe2x80x2) takes a maximum value under the condition of rxe2x80x2=rxc2x7sin xcex8 and xcfx86xe2x80x2=xcfx86i. If the following   a  =            r      xe2x80x2        r  
is assumed and arbitrary real number a satisfying 0 less than axe2x89xa61 is applied to evaluate V(xcfx86xe2x80x2) to find the maximum value, xcex8, A0, and xcfx86I can be all derived respectively from; xcex8=sinxe2x88x921 a       A    o    =            (              peak        ⁢                  xe2x80x83                ⁢        of        ⁢                  xe2x80x83                ⁢        V        ⁢                  xe2x80x83                ⁢                  (                      φ            xe2x80x2                    )                    )        a  
xcfx86i=(xcfx86xe2x80x2 providing peak) Incidentally, A0 is a correction term for an evaluation amplitude with sin xcex8.
An exponential function term (exp term) in the expression (8), i.e. vibration term for integrand is expressed as follows;
ejk(r sin xcex8 cos(xcfx86+xcfx86xe2x88x92xcfx86i)xe2x88x92rxe2x80x2 cos xcfx86)
If r greater than  greater than xcex, the value is substantially zero since expression (8) is an integral for the vibration solution unless xe2x80x9crxe2x80x2=rxc2x7sin xcex8 and xcfx86xe2x80x2=xcfx86ixe2x80x9d. On the other hand, when xe2x80x9crxe2x80x2=rxc2x7sin xcex8 and xcfx86xe2x80x2=xcfx86ixe2x80x9d, non vibration solution e0=1 is obtained to provide a peak.
Here, the result of expression (8) verified by a computer simulation will be described.
FIG. 5 shows a graph illustrating the result of estimating expression (8) using an electric field observed when A0xc2x7sinxcex8=1 for each wave source and a total of four wave sources S are arranged at xcfx86i=45 degrees, 135 degrees, 225 degrees, and 315 degrees, respectively, with a horizontal axis representing angle xcfx86, and a vertical axis DOA (Direction of Arrival) evaluation value. In this graph, a solid line represents the result when xcex8 for all the wave sources is 45 degrees while a dotted line represents the result when xcex8 for each wave source is changed in a range from 90 degrees to 30 degrees. In evaluating expression (8), rxe2x80x2=rxc2x7sin (45 degrees). Also, r=100 cm and xcex=15 cm.
As is apparent from FIG. 5, directions of a plurality of wave sources can be efficiently and accurately isolated according to the method of the present invention. Even when incident angle xcex8 with respect to z axis is unknown, xcex8 can be estimated by changing rxe2x80x2 in expression (8) to consider the peak level and the extended spectrum with respect to angle xcfx86. In this event, it is effective, for example, that all the spectrums with respect to angle xcfx86 in expression (8) are once derived with rxe2x80x2=r, and a maximum evaluation value is derived within a range of 0 less than rxe2x80x2xe2x89xa6r only for some peaks, from the fact that the peak position is not changed irrespective of xcex8. The spectrum here refers to a graph representing a change of the evaluation value with respect to rotation angle xcfx86.
Additionally, it is also possible to evaluate the component of rotation angle xcfx86 for electric field E, i.e. Excfx86. When the component of
Excfx86(r,xcfx86)
on the circumference with radius r in plane z=0 is assumed as follows;
Excfx86(r,xcfx86)=Aoxc2x7cos xcex8ejkr sin xcex8 cos xcfx86xe2x80x83xe2x80x83(9)
and considered as follows;
W(xcfx86)={fraction (1/xcfx80)}(1+cos(2xcfx86))xc2x7cos xcfx86xe2x80x83xe2x80x83(10)
then, it can be treated similarly to expression (5). In this case, however, the correction for the evaluation amplitude with sin xcex8 is not required.
Next, a multi-dimensional hologram data processing apparatus according to the present invention has an N-dimensional data array memory for writing an N-dimensional array data therein; an N-dimensional flag array memory for writing an N-dimensional array flag value therein; an N-dimensional array address generating unit for generating addresses for the N-dimensional array data; and a digital signal processing unit for controlling the N-dimensional array address generating unit, executing an algorithm, and outputting a peak point detection value and area detection value.
A method for extracting a plurality of peak points for multi-dimensional hologram data and an area occupied by the peak points using the multi-dimensional hologram data processing apparatus according to the present invention has:
a first step of setting all contents of the N-dimensional flag array memory to zero using the N-dimensional array address generating unit;
a second step of specifying a value of the N-dimensional array data providing a maximum value and an address value of the N-dimensional array data providing the maximum value in addresses in which the flag value represents zero using the N-dimensional array address generating unit, and proceeding to next step when one or more addresses in which the flag value represents zero are present and terminating the processing when no address representing a flag value of zero is present;
a third step of setting the flag value to zero and setting the content of the N-dimensional flag array memory of the above-mentioned address value to zero, and outputting the value of the N-dimensional array data providing the maximum value and the address value of the N-dimensional array data providing the maximum value as a peak point detection value;
a fourth step of setting a counter to zero;
a fifth step of specifying an address value in which the content of the N-dimensional flag array memory coincides with the flag value for all the addresses using the N-dimensional array address generating unit;
a sixth step of generating all address values tangent to a circumferential direction centered on the address value specified in the fifth step, adding 1 to the counter when the content of the N-dimensional flag array memory for the generated address value coincides with zero and the value of the N-dimensional array data value for the generated address value is smaller than the N-dimensional array data value for the address value specified in the fifth step, and rewriting the content of the N-dimensional flag array memory for the generated address value into the value obtained by adding 1 to the flag value;
a seventh step of determining whether the counter is equal to zero or not, and when not, adding 1 to the flag value for a new flag value and returning to the fourth step when the counter is not equal to zero, and proceeding to the next step when the counter is equal to zero; and
an eighth step of rewriting the content of the N-dimensional flag array memory into a maximum value representable as a flag value for all the addresses and outputting the address value thus obtained as an area address using the N-dimensional array address generating unit when the content of the N-dimensional flag array memory is other than zero and does not coincide with the maximum value representable as a flag value,
returning to the second step from the eighth step to repeat each step thereafter.
A hysteresis level may be set in the sixth step of generating all address values tangent to the circumferential direction centered on the address value specified in the fifth step.
The N-dimensional array data may be represented in a spherical coordinate system or a cylindrical coordinate system, and the end continuity may be maintained at the sixth step of generating all the address values tangent to the circumferential direction centered on the address value specified in the fifth step.
Therefore, the multi-dimensional hologram data processing apparatus and the method using the same for extracting a plurality of peak points for the multi-dimensional hologram data and the area occupied by the peak points can be used to facilitate the implementation of hardware and digital signal processing (DSP) with a simple algorithm, and achieve the processing at a speed hundred times faster than the prior art.